A visual inspection method and system for motor production flow

By constructing electrostatic interference feature representation and feature matching, the problem of misjudgment caused by electrostatic interference in motor production was solved, realizing the efficient and accurate identification of defects on the surface of insulating components by the vision inspection system, and improving production efficiency and the stability of quality control.

CN122199422APending Publication Date: 2026-06-12LUAN VOCATIONAL TECHNOLOGICAL COLLEGE
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
LUAN VOCATIONAL TECHNOLOGICAL COLLEGE
Filing Date
2026-03-04
Publication Date
2026-06-12

AI Technical Summary

Technical Problem

During the motor manufacturing process, image misjudgment caused by electrostatic interference means that existing vision inspection systems cannot effectively distinguish between electrostatic interference and real defects, leading to false alarms and reduced production efficiency.

Method used

By constructing a feature representation of electrostatic interference, utilizing the brightness dispersion and gradient direction consistency features of image patches, and combining a pre-trained classifier and defect feature templates, electrostatic interference can be distinguished from real defects. Feature matching and continuous image analysis are employed to ensure the accuracy of the detection results.

Benefits of technology

It significantly reduces the probability of misjudgment caused by electrostatic interference, improves the reliability and consistency of test results, and ensures the stability of quality monitoring of insulating components and the controllability of the production process.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122199422A_ABST
    Figure CN122199422A_ABST
Patent Text Reader

Abstract

The application relates to the technical field of computer vision, and particularly discloses a visual detection method and system for motor production processes, which comprises the following steps: continuously acquiring n surface images of insulation components above a conveying belt of a production line by an imaging unit, and pre-processing the current surface image to obtain an image to be analyzed; constructing an electrostatic interference feature expression of the image to be analyzed, the electrostatic interference feature expression being formed based on a rapid spatial change mode of pixel brightness in the image to be analyzed, and the rapid spatial change mode being embodied as a non-directional sharp fluctuation of the pixel brightness in a region; determining the abnormal source in the current surface image, the abnormal source including electrostatic interference and real defects, and if the abnormal sources of the continuous n / 5 surface images are all real defects, determining that the insulation component is unqualified. The application can improve the consistency and accuracy of defect determination.
Need to check novelty before this filing date? Find Prior Art